Meso- and bathypelagic distribution and abundance of chaetognaths in the Atlantic sector of the Southern Ocean

(1)

DOI 10.1007/s00300-009-0632-3 O R I G I N A L P A P E R

Meso- and bathypelagic distribution and abundance

of chaetognaths in the Atlantic sector of the Southern Ocean

Svenja Kruse · Ulrich Bathmann · Thomas Brey

Received: 5 December 2008 / Revised: 3 April 2009 / Accepted: 3 April 2009 / Published online: 19 April 2009

Abstract We conducted multinet sampling during winter and summer in the Southern Ocean (Atlantic sector) to investi- gate the eVect of water mass, season and water depth on abundance and species composition of meso- and bathypelagic chaetognaths. Eukrohnia hamata (mean 115 ind. 1,000 m^¡3) and Sagitta marri (mean 51 ind. 1,000 m^¡3) were dominant, complemented by E. bathypelagica (mean 19 ind. 1,000 m^¡3) and E. bathyantarctica (mean 19 ind. 1,000 m^¡3) below 1,000 m. A further six species were identiWed, among them the rare bathypelagic species Heterokrohnia fragilis and the sub- tropical Eukrohnia macroneura that is new to the Antarctic.

Water depth and season were the principal determinants of abundance and species composition patterns, indicating vertical seasonal migration and vertical segregation of species. The life cycles of E. hamata and S. marri were studied additionally.

Their maturity stages were vertically segregated and prolonged reproductive periods are suggested for both species.

Keywords Chaetognatha · Antarctica · Bathypelagial · Distribution · Abundance · Life cycle

Introduction

Chaetognaths represent a major component of the world’s marine zooplankton. In the Southern Ocean they contribute

signiWcantly to the total zooplankton stock, at times reach- ing up to 30% of the total zooplankton abundance (Piatkowski 1985; Froneman and Pakhomov 1998; Pakhomov et al. 1999, 2000). As main predators of copepods (Øresland 1990, 1995) chaetognaths may consume up to 5.2% of the standing stock per day (Froneman and Pakhomov 1998). Hence, they are of great importance for the energy transfer from copepods to higher trophic levels (Bone et al.

1991) and may contribute considerably to the vertical carbon Xux (Dilling and Alldredge 1993).

Detailed studies on the Antarctic chaetognath fauna started at the beginning of the twentieth century (e.g. by Ritter-Záhony 1911), already more than 100 years after the Wrst publication concerning a chaetognath (Slabber 1775, reviewed by Bone et al. 1991). So far, investigations on Antarctic chaetognath ecology focused on the austral summer and on the upper 500 m (e.g. Timonin 1968; Terazaki 1989; Bielecka and Zmijewska 1993; Blachowiak-Samolyk et al. 1995) to 1,000 m (Thiel 1938; Duró et al. 1999; Duró and Gili 2001; Johnson and Terazaki 2004) of the water column. Despite the extensive data on Antarctic chaetognath distribution and abundance below 1,000 m of David (1958a, 1965) and Alvariño et al. (1983a, b), our knowl- edge of the deep water chaetognath ecology is still frag- mentary. There is a general lack of deep samples, and, quite often, unsuitable large mesh sizes were used (Hagen 1985;

Duró and Gili 2001). Consequently, reliable quantitative data are rare, and hitherto a number of bathypelagic species are known from very few specimens only (Terazaki 1991).

One major objective of our study was to evaluate the eVects of water mass (Polar Frontal Zone, Weddell Gyre, Coastal Current), of season (summer–winter) and of water depth (4 depth strata) on abundance and species composition of meso- and bathypelagic chaetognaths in the Atlantic sector of the Southern Ocean. Furthermore, the two expeditions S. Kruse (&) · U. Bathmann · T. Brey

Alfred Wegener Institute for Polar and Marine Research, Am Handelshafen 12, 27570 Bremerhaven, Germany e-mail: Svenja.Kruse@awi.de

U. Bathmann

e-mail: Ulrich.Bathmann@awi.de T. Brey

e-mail: Thomas.Brey@awi.de

(2)

provided a unique opportunity for seasonal deep sampling with small mesh sizes (100m) in the same area, thus allowing an investigation of the complete community composition covering the entire size range and all maturity stages of the predominant chaetognath species. Detailed life cycle analyses were possible, and contribute to our knowl- edge on chaetognath biology in the Southern Ocean.

Materials and methods

Field sampling

Chaetognaths were sampled during two expeditions in the Lazarev Sea with the RV Polarstern, expedition ANT 23-6 in Antarctic winter 2006 (17 June–21. August 2006), and expedition ANT 24-2 in Antarctic summer 2007/2008 (28 November 2007–04 February 2008). StratiWed sampling with a multinet was performed at 28 stations in winter (between 60° and 68°30⬘S) and at 15 stations in summer (at 52°S and between 62° and 70°S) along three transects (3°W, 3°E and 0°E). This multiple opening/closing net (opening size: 0.25 m²) was equipped with Wve nets with 100-m mesh size and sampled the following standard depth intervals: 2,000–1,500, 1,500–1,000, 1,000–750, 750–500, 500–0 m. Exceptions from the standard depths were made at three stations during ANT 23-6 (at 61°30⬘S and 62°S 3°E to 3,000 m, at 65°S 3°E to 1,250 m depth) and at one during ANT 24-2 (at 70°S 3°W to 1,500 m depth). The winter station around 66°S 0°E was a 5-days station, located at a drifting ice camp.

As the abundance of chaetognaths in the epipelagial is already well known and as we are particularly interested in meso- and bathypelagic chaetognaths, we neglected the 500–0 m depth layer in the present study.

Our sampling scheme covered three diVerent water masses, the Polar Frontal Zone (PFZ) with two stations at 52°S in summer, the Weddell Gyre (WG), water mass between 60°S and 68°S, and the Coastal Current (CC) at and south of 68°S. The diVerent pelagic zones are deWned as follows: epipelagic (0–500 m), mesopelagic (500–

1,000 m) and bathypelagic (below 1,000 m).

Laboratory methods and data processing

Directly after sampling, chaetognaths were sorted. The specimens were counted, identiWed to species level and their body length (without tail Wn) was measured under a stereomicroscope (Olympus SZX12) to the nearest 0.5 mm.

During the winter expedition, a part of each sample was immediately preserved in formaldehyde (4% Wnal concentration, buVered with hexamine) and measured later in the home laboratory. To compensate for preservation induced shrinkage, we computed shrinkage factors for the dominant species from repeated length measurements of fresh and subsequently formaldehyde preserved specimens collected during the summer expedition. This allowed the comparison of lengths between formaldehyde preserved and frozen chaetognaths.

Taxonomic identiWcation was conducted to species level under a stereomicroscope (see above) and a microscope (Zeiss Axioskop 2 plus) using the relevant literature (Alvariño 1969; O’Sullivan 1982; Casanova 1986, 1999;

Kapp 1991a). Damaged chaetognaths, that could not be identiWed to species level, or smaller Eukrohnia individuals (<10 mm) were pooled as Sagitta or Eukrohnia spp., respectively. The two most abundant species, Eukrohnia hamata Möbius 1875 and Sagitta marri David 1956, were classiWed into Wve maturity stages according to Kramp (1939) and David (1955) (Table1).

Table 1 Maturity stage classiWcation of Eukrohnia hamata and Sagitta marri according to Kramp (1939) and David (1955) Stage Eukrohnia hamata (from Kramp 1939) Sagitta marri (from David 1955 for Sagitta gazellae)

Male gonads Female gonads Male gonads Female gonads

I Unripe Unripe Tail segment empty;

rudiments of testes present

Ovaries not visible or rudimentary II Tail containing more

or less sperm

All eggs small Tail segment opaque;

seminal vesicles may show as small protuberances

Ovaries short and thin; eggs small

III Sperm evacuated All eggs small, seminal receptacles Wlled with sperm

Seminal vesicles fully formed;

tail segment empty

Ovaries thin, but variable in length

IV Sperm evacuated Ovaries Wlled with ripe eggs Seminal vesicles usually discharged

Ovaries thick and long; eggs enlarged V Sperm evacuated Eggs evacuated, receptacles

still containing sperm

Sperm discharged Eggs discharged; remnants of ovaries are irregular masses sometimes spread into the tail segment

(3)

In the genus Eukrohnia, we pooled all individuals smaller than 6 mm (and usually larger than 2.5 mm) belonging to stage 1 into the group “Eukrohnia juveniles”.

The small individuals of Eukrohnia bathyantarctica David 1958 could easily be identiWed, but the juveniles of E. hamata and of E. bathypelagica Alvariño 1962 were diYcult to distinguish, owing to lack of characters and con- gruence of size. SpeciWc characteristics of adults such as a Xabby, translucent body or coiled immature ovaries and a proportionally longer tail (described by Alvariño 1962) are not yet developed in juveniles of 5 mm length and this results in problems of species identiWcation. Because of the absence of stages 4 and 5 individuals of E. hamata in summer, we suggest that the remaining juveniles are E. bathypelagica, as stage 4 and 5 individuals of this species were observed (Kruse 2009). It is possible, however, that we just missed the mature E. hamata adults (as discussed below).

Numbers per sample are standardized to number of individuals per 1,000 m³. For the 5-days winter station the geographical and vertical abundance data are averaged over all eight sub-stations.

Statistical analyses

To evaluate diVerences in species composition, we applied a cluster analysis (e.g. Everitt et al. 2001) to the species£abundance matrix (9 species£170 samples, see Sect. “Results”). The resulting cluster identities were taken as representative for distinct species assemblages. Nomi- nal logistic regression (e.g. Agresti 2002) was used to identify relationships between cluster identity and water mass, season and depth layer. Abundance values were square-root transformed prior to analysis to reduce the inXuence of outliers. We applied hierarchical clustering and compared several linkage methods (average, centroid, complete, Ward’s minimum variance) to check for consis- tency of results.

We analysed abundance data at the family level (Sagit- toidea, i.e. all species present), at the genus level (Eukroh- nia and Sagitta) and at the species level (abundant species only, see below). Maturity stages (mean stage per sample) were analysed for E. hamata and S. marri. Data were Box- Cox transformed to achieve normality and homogeneity of variances and subjected to a full factorial three-way ANOVA (abundance/mean stage versus water mass and season and depth and water mass£depth and season£depth) with subsequent post hoc test on diVer- ences between means (= 0.05, Sokal and Rohlf 1981).

The interaction term water mass£season was not tested, as there are no winter samples from the PFZ.

Additionally, a full factorial two-way ANOVA (length versus maturity stage and season and maturity

stage£season) was applied to analyse diVerences in length in E. hamata and S. marri (data were treated as mentioned for the previous ANOVA). Seasonal diVerences between the length–frequency distributions were analysed by means of a Kolmogorow–Smirnow test in both species.

All statistical analyses were performed with the software package JMP (SAS Inc).

Results

InXuence of formaldehyde on chaetognath body length Due to the preservation of the samples with formaldehyde (4% Wnal concentration, buVered with hexamine, 4 months exposure) the chaetognath body length shrunk up to 21%.

Shrinkage amounted to 3.67% (SD§2.51, n= 104) in Eukrohnia hamata, to 5.37% (§3.38, n= 93) in E. bathy- antarctica, and to 6.23% (§3.84, n= 79) in E. bathypelag- ica irrespective of length and maturity stage. Highest reduction of 7.17% (§3.97, n= 87) in length was measured for S. marri. The chaetognaths shrunk particularly in the Wrst days and weeks. However, they kept shrinking very slowly even after 4 months of formaldehyde preservation (personal observation).

Geographical and vertical chaetognath distribution

We were able to identify ten diVerent species from three genera in our samples: E. hamata, E. bathypelagica, E. bathyantarctica, E. macroneura Casanova 1986, Heterokrohnia fragilis Kapp and Hagen 1985, H. mirabilis Ritter-Záhony 1911, S. marri, S. macrocephala Fowler 1905, S. maxima Conant 1896 and S. gazellae Ritter- Záhony 1909. E. hamata and S. marri were the two most abundant of these species, independent of the water masses (Tables2, 3, 4).

Eukrohnia juveniles were very frequent in summer, and for a better comparison of seasons they were excluded from the Figs.1b, 2b, 3b, 4b that display summer data but are presented separately (Fig.1c). During winter Eukrohnia juveniles were extremely rare and thus are not presented separately. Juveniles of other species, e.g. Sagitta marri, were readily identiWed and not treated separately.

The mean chaetognath abundance of the 500 to 2,000 m depth stratum ranged from 58 ind. 1,000 m^¡³ (61°30⬘S 3°E) to 443 ind. 1,000 m^¡³ (65°S 3°E; Fig.1a) in winter, and from 91 ind. 1,000 m^¡³ (64°30⬘S 0°E) to 508 ind. 1,000 m^¡³ (70°S 3°W) in summer (without Eukrohnia juveniles; Fig.1b). Juvenile Eukrohnia ranged from 2 ind. 1,000 m^¡³ (69°S 0°E) to 880 ind. 1,000 m^¡³ (66°S 3°E, 62°S 0°E, Fig.1c) in summer.

(4)

Table2Chaetognath species abundance (individuals1,000m¡3 ) and relative composition for each depth interval, presented for the summer and winter situation in CC (Coastal Current) n the number of investigated stations CCSummer (n=4)Winter (n=4) 500–750m750–1,000m1,000–1,500m1,500–2,000m500–750m750–1,000m1,000–1,500m1,500–2,000m Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD% Eukrohnia bathyantarctica000471.0411722.7492446.628385.2893.1301411.7563527.2 Eukrohnia bathypelagica21271.88162.41347.5271725.212242.2481.6301411.7241111.6 Eukrohnia hamata42525835.2782223.110135.8151314.43402462.51489857.81206346.9847340.8 Eukrohnia macroneura00019155.816129.100000012244.7481.6000 Eukrohnia spp.48264.08172.59115.110199.0480.7366214.0424316.4241.0 Juvenile Eukrohnia53959944.716830849.7331518.6000000000000000 Heterokrohnia fragilis000000000000000000000000 Heterokrohnia mirabilis000000000000000000000000 Sagitta gazellae000000000000000481.6000451.9 Sagitta macrocephala000000000000000000000000 Sagitta marri17212314.3526515.5549330.10001604129.4403115.6262010.1363017.5 Sagitta maxima000000000342.4000481.6451.6000 Sagitta spp.000000000000000000000000 Juvenile chaetognaths000000241.1342.4000000000000 UnidentiWed000000000000000000000000

(5)

Table3Chaetognath species abundance (individuals1,000m¡3) and relative composition for each depth interval, presented for the summer and winter situation in WG (Weddell Gyre) n the number of investigated stations WGSummer (n=9)Winter (n=24)Winter (n=2) 500–750m750–1,000m1,000–1,500m1,500–2,000m500–750m750–1,000m1,000–1,500m1,500–2,000m2,000–3,000m Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD%Mean§SD% E. bathyantarctica12300.59180.8472433.7401934.0270.54141.312145.6463431.318856.25 E. bathypelagica42311.918201.5161211.4331727.911312.68162.5261912.0292319.76318.75 E. hamata47424221.295618.0171911.9131511.327816267.114611846.1876239.9484032.5236.25 E. macroneura00015221.2462.60000003111.1230.80.320.2000 Eukrohnia spp.1251035.67120.6151510.710118.917304.223307.210114.4372.3236.25 JuvenileEukrohnia1,3301,36359.392092377.7253217.4130.80000.730.25122.13152.4000 H. fragilis000000000685.2000000000000000 H. mirabilis000000000000000000000000236.25 S. gazellae7110.3250.1261.3463.09152.2150.4240.8150.9000 S. macrocephala000000000000000000000000000 S. marri25211611.2107709.09146.4363.0968823.21308941.0747133.8152310.0236.25 S. maxima0000000000000.730.20.730.20.320.10.30.20.2000 Sagitta spp.000250.2130.6241.5000000000000000 Juvenile chaetognaths0007160.66124.0000000000000000000 UnidentiWed0004110.30005164.4000000140.5130.5000

(6)

Regarding water depth, highest abundances were encountered between 500 and 1,000 m in winter (Figs.2a, 3a, 4a), attaining values up to 1,248 ind. 1,000 m^¡³ (500–750 m, 64°S 0°E, Fig.3a), and between 500 to 750 m in summer (Figs.2b, 3b, 4b), with a maximum of 1,470 ind. 1,000 m^¡³ (63°S 3°E, Fig.4b). Eukrohnia juveniles did rarely occur deeper than 1,500 m and distinctly preferred the 500–

1,000 m depth range in summer (Tables2, 3, 4).

Chaetognath species composition

The cluster analysis of the 9 species£170 samples matrix (we excluded Heterokrohnia mirabilis, because it occurred in one of the two exceptional samples collected below 2,000 m only) produced a rather consistent sample group- ing pattern, irrespective of the linkage method applied. Spe- cies composition was signiWcantly aVected by water depth (P< 0.001, ²= 144.78), season (P< 0.001, ²= 45.65) and water mass (P= 0.001, ²= 32.98; eVect likelihood ratio test of the nominal logistic regression). The eVect of water depth was mainly related to E. bathyantarctica and E.

bathypelagica which dominated the deeper community but were almost absent in the upper layers, and to E. hamata that showed the opposite pattern (Tables2, 3, 4). The seasonal eVect was related to the less frequent species.

E. bathyantarctica, E. bathypelagica, E. macroneura and H. fragilis were more frequent in summer, whereas S. gazellae was more frequent in winter. The water mass

eVect was most likely caused by the (non-) occurrence of species in just one water mass, such as Sagitta macrocep- hala and H. fragilis that occurred exclusively in the PFZ and the WG, respectively.

Chaetognath abundance

ANOVA of abundance data at the genus and the species level indicated that water mass had barely any eVect, only the abundance of E. bathypelagica was signiWcantly higher in Polar Frontal Zone (PFZ) than in Weddell Gyre (WG) and Coastal Current (CC, Table5). The same holds true for the interaction of water mass and depth. Here, we found a signiWcant eVect on all species pooled (class Sagittoidea), where abundance decreased with depth within the WG and CC, and in the genus Eukrohnia, where it decreased only within the WG.

SigniWcant seasonal diVerences were detected in the genera Eukrohnia and Sagitta (Table5). Sagitta was more abundant in winter than in summer and Eukrohnia vice versa. Within the genus Sagitta, S. marri was 8 times more abundant in the 1,000–1,500 m stratum in winter (74 ind. 1,000 m^¡³ in WG, Table3) than in summer. The higher abundance of Eukrohnia in summer, however, can be attributed to the high number of juveniles, as the dominant E. hamata was again signiWcantly more abundant in winter.

Depth had the most distinct eVect on chaetognath abundance. E. hamata, the dominant species, was signiWcantly more abundant in the 500–750 m depth range than at Table 4 Chaetognath species abundance (individuals 1,000 m^¡3) and relative composition for each depth interval, presented for the summer and winter situation in PFZ (Polar Frontal Zone)

n the number of investigated stations

PFZ Summer (n= 2)

500–750 m 750–1,000 m 1,000–1,500 m 1,500–2,000 m

Mean §SD % Mean §SD % Mean §SD % Mean §SD %

Eukrohnia bathyantarctica 0 0 0 32 22 8.2 16 11 9.1 48 35 35.7

Eukrohnia bathypelagica 47 20 16.3 40 34 10.2 89 11 50.0 32 22 23.4

Eukrohnia hamata 147 92 51.4 79 22 20.4 32 0 18.2 16 0 11.8

Eukrohnia macroneura 0 0 0 8 11 2.0 28 6 15.9 0 0 0

Eukrohnia spp. 15 21 5.3 32 0 8.2 0 0 0 4 6 2.9

Juvenile Eukrohnia 23 32 8.0 143 202 36.8 0 0 0 24 33 17.4

Heterokrohnia fragilis 0 0 0 0 0 0 0 0 0 0 0 0

Heterokrohnia mirabilis 0 0 0 0 0 0 0 0 0 0 0 0

Sagitta gazellae 0 0 0 0 0 0 0 0 0 0 0 0

Sagitta macrocephala 0 0 0 8 11 2.0 0 0 0 4 6 2.9

Sagitta marri 46 42 16.1 32 22 8.2 4 6 2.2 0 0 0

Sagitta maxima 8 12 2.9 8 11 2.0 8 11 4.6 4 6 2.9

Sagitta spp. 0 0 0 8 11 2.0 0 0 0 0 0 0

Juvenile chaetognaths 0 0 0 0 0 0 0 0 0 0 0 0

UnidentiWed 0 0 0 0 0 0 0 0 0 4 6 3.0

(7)

greater depths (Table5), attaining maximum summer abundances of 425 and 474 ind. 1,000 m^¡³ in the CC and WG, respectively. Sagitta marri, which was second in abundance, preferred a wider depth range, 500 and

1,000 m, with a maximum of 252 ind. 1,000 m^¡³ in the WG in summer (500–750 m, Table3). Eukrohnia bathype- lagica and E. bathyantarctica showed the opposite abun- dance pattern, as they preferred layers below 1,000 m Fig. 1 Geographical distribu-

tion and mean abundance of chaetognaths along the three sampling transects during winter (a) and summer (b) without juveniles. The juveniles from the summer expedition are presented separately (c). PFZ Polar Frontal Zone, WG Weddell Gyre, CC Coastal Current.

Stations at and south of 68°S are considered within the CC (horizontal line)

(8)

(Table5). They reached highest numbers of 89 ind.

1,000 m^¡³ (summer, PFZ, 1,000–1,500 m, Table4) and 56 ind. 1,000 m^¡³ (winter, CC, 1,500–2,000 m, Table2), respectively. Eukrohnia macroneura diVered from all other species, as it was most abundant between 750 and 1,500 m depth (maximum of 28 ind. 1,000 m^¡³, summer, PFZ,

Table4). Although juveniles of the genus Eukrohnia could not be subjected to sound statistical analyses, their centre of abundance was observed between 500 and 1,000 m, with values up to 1,330 ind. 1,000 m^¡³ in the WG (500-750 m, Table3), then representing 59.3% of total chaetognath abundance.

Fig. 2 Vertical distribution and abundance of chaetognaths (without juveniles) along the 3°W transect during winter (a) and summer (b). WG Weddell Gyre, CC Coastal Current.

Stations at and south of 68°S are considered within the CC (vertical line)

Fig. 3 Vertical distribution and abundance of chaetognaths (without juveniles) along the prime meridian during winter (a) and summer (b). WG Weddell Gyre, CC Coastal Current.

(9)

In species that showed decreasing abundance with depth (E. hamata, S. marri), this vertical gradient became signiW- cantly more distinct in summer, as indicated by the season£depth interaction term of the ANOVA (Table5).

In contrast, E. bathypelagica exhibited a more distinct depth gradient in winter, with a clear preference for the 1,000–2,000 m layer which, however, was caused mainly by a decrease in abundance in shallower layers from summer to winter.

In Sagitta gazellae we could not detect any eVects of water mass, season or depth. All other species were too rare for reliable analysis. S. macrocephala was only captured in summer in the PFZ (8 ind. 1,000 m^¡³ in 750–1,000 m, Table4). S. maxima was primarily found in the PFZ as well, where this species was encountered between 500 and 1,500 m in summer (8 ind. 1,000 m^¡³, Table4). Two species of Heterokrohnia were found below 1,500 m in the WG (Table3). A total number of 7 individuals of H. fragi- lis (7–10 mm length) were caught in summer, H. mirabilis occurred exclusively between 2,000 and 3,000 m in winter and with 2 ind. 1,000 m^¡³ (6.3%) was even rarer than H.

fragilis. Only one H. mirabilis specimen of 19 mm length was caught in the WG.

Distribution of maturity stages in E. hamata and S. marri ANOVA of mean maturity stage showed a signiWcant eVect of all parameters investigated (Table5). In both species, mean maturity stage was higher in the PFZ than in the WG.

Season aVected E. hamata and S. marri diVerently; the former species showed higher mean maturity in winter, the latter in summer. Generally, mean maturity stage increased with depth. However, in E. hamata no signiWcant diVer- ences were detected below 750 m. In S. marri mean maturity stage was signiWcantly higher in the 1,500–2,000 m stratum compared to the 750–1,000 m stratum. The interaction of season and depth indicated that in E. hamata the vertical gradient was more distinct in winter, in S. marri however, in summer, as in this species depth had no eVect at all in winter.

Population structure of E. hamata and S. marri Eukrohnia hamata

Of all Eukrohnia hamata caught, 99.6% (summer) and 99.9% (winter) were complete and could be measured. The population of E. hamata consisted essentially of stages 1 and 2 individuals (Fig.5). E. hamata had a maximum length of 29 mm in summer and 32 mm in winter, respectively (Table6). During both seasons their length increased slightly with increasing depth, as maturity stage and body length are positively correlated (winter: r= 0.764, P< 0.001; summer: r= 0.813, P< 0.001), albeit with much overlap in length between subsequent stages (Fig.5). Com- paring both seasons the mean body length per stage did not diVer signiWcantly between seasons: stage 1:14.6 mm, stage 2:23.3 mm, and stage 3:27.7 mm.

Fig. 4 Vertical distribution and abundance of chaetognaths (without juveniles) along the 3°E transect during winter (a) and summer (b). WG Weddell Gyre, CC Coastal Current.

(10)

The shape of the length–frequency distribution diVered signiWcantly between summer and winter (Kolmogorow–

Smirnow test, P< 0.005). Apparently there was a higher proportion of large animals (>20 mm) present in winter.

This coincides with a signiWcantly higher mean maturity

stage in winter (see above). Including the unidentiWed Eukrohnia individuals would slightly increase the stage 1 individuals (especially below 10 mm length), but not sig- niWcantly change the size–frequency structure (Kolmogorow–

Smirnow test, P> 0.1).

Table 5 EVects of water mass WM, season and depth on chaetognath abundance and maturity stage distribution (mean maturity stage per station and depth interval)

Full factorial (except WM£season) ANOVA with subsequent Tukey HSD post hoc test on diVerences between means (= 0.05), letters (A, B…) indicate groups that diVer signiWcantly, the alphabetical order indicates decreasing abundance/mean maturity stage. The interaction term WM£depth is not shown here, because it was signiWcant only for the class Sagittoidea where abundance decreased with depth in WG and CC, but not in PFZ, and for the genus Eukrohnia where abundance decreased with depth within WG. Sagitta gazellae is not mentioned in this table, because all tests were not signiWcant

PFZ Polar Frontal Zone, WG Weddell Gyre, CC Coastal Current, S summer, W winter, 1: 2,000–1,500 m, 2: 1,500–1,000 m, 3: 1,000–750 m, 4:

750–500 m), ns no signiWcant eVect

Water mass Season Depth Season£depth

PFZ WG CC W S 1 2 3 4 W 1 W 2 W 3 W 4 S 1 S 2 S 3 S 4

Abundance

Class Sagittoidea ns A A A A A

B B B B B B B B

C C C C C C

Genus Eukrohnia ns A A A A A

B B B B B B B B B

C C C C C C

Genus Sagitta ns A A A A A A A

B B B B B B

C C C C

E. bathyantarctica ns A A A A A A

B B B B B

C C C C

E. bathypelagica A A A A A A A A A A

B B B B B B B B

E. hamata ns A A A A

B B B B B

C C C C C

D D

E. macroneura ns A A A A A

B B B B B B B B B

S. marri ns A A A A A A A

B B B B B B

C C C C

Mean stage

E. hamata A A A A A A A A A A

B B B B B B B B B B

C C

S. marri A A A A A A

B B B B B B B B B B B B

C C

(11)

Sagitta marri

All specimens of Sagitta marri could be measured in summer, during winter the measuring success rate was 93.3%.

This population was dominated by maturity stages 1 and 2 (Fig.6). Maximum body length was 27 mm in summer and 28 mm in winter, respectively (Table7). The stage-speciWc mean lengths diVered seasonally only between stage 1 specimens. In the 500–750 m layer S. marri had, e.g. a mean length of 6.8 mm in summer compared to 9.4 mm in winter. Maturity stage and length were positively correlated

in this species, too (winter: r= 0.636, P< 0.001; summer:

r= 0.801, P< 0.001).

The population size–frequency structure (Fig.6) did not diVer signiWcantly between winter and summer (Kolmogorow–Smirnow test, P> 0.1). As a result of longer specimens (see above), the structure for the winter situation was shifted towards greater lengths with highest values between 8 and 11 mm body length dominated by stage 2 individuals. Highest values in summer were shown at 6 and 7 mm body length represented by stage 1 individuals.

Fig. 5 Length–frequency and corresponding maturity stage distribution of Eukrohnia hamata in winter and summer.

n the number of investigated individuals

winter n=1253

stage composition [%]stage composition [%]

body length [mm]

summer n=550

(12)

Discussion

Of the ten species found during this investigation, E.

hamata, S. gazellae, S. marri and S. maxima were most fre- quently recorded in the past 50 years of Antarctic research (David 1958a; Alvariño 1969; Dinofrio 1973; Alvariño et al. 1983a, b; Hagen 1985; Johnson and Terazaki 2004).

DiVerent nets of varying and rather coarse mesh sizes were used in previous studies compared to our multinet with 100-m mesh size. Thus, abundance data are diYcult to compare, as we have caught smaller (younger) animals with higher eYciency, but larger chaetognaths (e.g. S. gazellae, Hagen 1985) may be underrepresented to some extend due to active escape reactions.

Parameters inXuencing chaetognath abundance and species composition

Water mass

Spatial variability of chaetognath abundance is enormous (Fig.1), even on small scales, as observed during all hauls at the station located at the ice camp (WG) within 5 days and 32 nm total drift distance (start to end distance, 7 nm).

This patchiness, that is typical for zooplankton, might have obscured to a large extent diVerences in chaetognath abundance and composition between the three diVerent water masses PFZ, WG and CC.

The sole Wnding of Sagitta macrocephala in the PFZ supports previous reports, as this species was described to be more frequent in the deep mesopelagic layers of the Sub- antarctic than in those of the Antarctic waters (David 1958a, 1965). We found just one signiWcant eVect of water mass: Eukrohnia bathypelagica was more abundant in the PFZ than in the other two water masses, particularly below 750 m (Tables2, 3, 4). At this depth a tongue of warm

(about 2°C) and saline (>34.7) water stretches from Subant- arctic into polar regions (Schröder and Fahrbach 1999).

One should keep in mind that our stations were situated at the southernmost edge of the PFZ or Antarctic Conver- gence. Thus, our data might not have caught the full impact of the particular PFZ hydrodynamics on chaetognath distribution. This might also explain to some extent that we did not see higher abundance of E. hamata in PFZ waters.

E. hamata, a cosmopolitan species (Alvariño 1969), is con- sidered to be the most abundant species in Subantarctic and Antarctic waters, showing maximum abundance in the vicinity of the Antarctic convergence where higher densities have been reported even deeper in the water column (David 1958a). In the top 500 m E. hamata is known to reach maximum concentrations (David 1958a, 1965; Johnson and Terazaki 2004); hence, generally higher densities of E. hamata may occur around the Antarctic Convergence in the epipelagic realm. In general, the upper layer of this water body reveals higher plankton concentrations than adjacent areas (Voronina 1968). Between 49° and 50°S, highest downward velocity is observed at 20°E which results in an increased zooplankton abundance especially in the upper 100 m (Voronina 1968). In the meso- and bathypelagial of the Antarctic Convergence, however, these diVerences in zooplankton density are probably not detect- able any more.

Water depth

Depth was found to be the major determinant of chaetognath abundance and distribution on all taxonomic levels.

Eukrohnia hamata was the dominant species in terms of abundance, especially between 500 and 1,000 m. It showed a signiWcant decrease in abundance with depth, a pattern already found in previous studies (e.g. Alvariño et al.

1983a, b). E. bathypelagica and E. bathyantarctica coexist Table 6 Length–frequency distribution for Eukrohnia hamata in the diVerent depth intervals for winter and summer

n the number of investigated individuals Depth (m) n Length (mm)

6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Winter

500–750 490 5 4 5 8 14 11 15 14 14 24 28 44 32 42 38 38 32 23 35 36 17 8 2 1

750–1,000 257 7 7 8 8 11 4 7 9 14 9 15 13 20 18 23 18 26 24 11 4 1

1,000–1,500 335 2 2 9 10 6 13 11 12 13 19 14 14 22 17 26 37 37 35 21 9 4 2

1,500–2,000 171 1 1 3 4 6 4 5 7 10 16 11 8 13 8 11 17 19 11 9 5 1 1

Summer

500–750 410 1 14 24 25 9 20 36 38 24 24 21 31 15 24 19 17 24 19 10 9 5 1

750–1,000 83 3 3 3 3 4 6 3 3 7 8 7 14 9 3 6 1

1,000–1,500 32 1 1 1 1 1 1 1 1 1 1 1 5 4 6 5 1

1,500–2,000 25 1 1 3 1 1 3 1 1 1 1 1 1 1 3 2 1 2

(13)

with, and partially displace, E. hamata in the deep meso- and bathypelagic oceanic strata. Whereas E. bathypelagica, a species with a worldwide distribution (e.g. Rottmann 1978, Gulf of Thailand; Terazaki 1996, Equatorial PaciWc), inhabits the layers below 500 m, E. bathyantarctica occurs mainly below 1,000 m in Antarctic waters.

Eukrohnia bathypelagica dominated the 1,000–2,000 m depth range and reached average numbers of up to 33 ind. 1,000 m^¡3 in summer in the WG (1,500-2,000 m) and 89 ind. 1,000 m^¡3 in the PFZ (1,000–1,500 m). So far, only Alvariño et al. (1983a, b) provided detailed informa- tion on the geographical and bathymetric distribution of

E. bathypelagica in the Southern Ocean. In summer, they observed low densities of E. bathypelagica (·10 ind.

1,000 m^¡3) in the meso- and bathypelagial of the Scotia Sea, Weddell Sea and the Drake Passage, which is three times less than we observed in average. The winter data given by Alvariño et al. (1983a) excluded the Weddell Sea, but data for the South PaciWc showed largest abundances of up to 1,000 ind. 1,000 m^¡3 below 1,000 m north of 60°S.

To the south this species always occurred with less than 100, in some areas dropped even below 10 ind. 1,000 m^¡3 in the meso- and bathypelagic zone. We note that Alvariño et al. (1983a) included the 200–500 m range in Fig. 6 Length–frequency and

corresponding maturity stage distribution of Sagitta marri in winter and summer. n the number of investigated individuals

stage composition [%]stage composition [%]

body length [mm]

summer n=315 winter n=640